Thursday, December 12, 2013

Introduction to SAP NetWeaver Business Warehouse

Why DW/BI System?

§  The goal behind the implementation of ERP systems and other software tools is the cost reduction, and automation of processes in individual business areas
§  The result is that these ERP systems, CRM systems, banking and credit card systems, etc. have exponentially increased data volumes needing analysis
§  Also, the ever-increasing globalization and, the increasing decentralization of organizations has created the need to recognize market trends and to collect information about competitors
§  Decision makers in modern, globally operating enterprises frequently realize that their survival depends on the effective use of this information
§  Unfortunately this information is often spread across many systems and sometimes many countries, thus making effective use of information extremely difficult
§  This is precisely the challenge that modern Business Intelligence systems attempt to meet

Evolution of SAP NetWeaver Business Warehouse

§  BI/BW is warehouse tool developed by SAP
§  First version was BW 1.2 (Business  Warehouse), 1998
§  BW 2.0 (Business  Warehouse), 2000
§  BW 3.x/3.5 (Business Warehouse), 2004
§  BI 7.x/7.0 (Business Intelegency), 2005
§  BW 7.3 (Business Warehouse), 2010
§  BW 7.3 (SP5)  (Business Warehouse ) HANA Enabled System, 2012
§  Up till version BI 7.x most BI systems used Oracle as a back end DBMS but with BW 7.3 (SP5) you have the option to deploy your BI system with SAP proprietary in-memory database HANA
§  So DW/BI/BW is one of the same thing
§  However, there is major changes b/w version 3.x and 7.x, specially the data flow has entirely changed b/w 3.x and 7.x
§  We still need to learn the data flow of 3.x, because SAP has pre-defined most of the contents called BI Content (InfoCubes, Transformation, DataSources etc.)

BI Contents
§  SAP has vast experience in business process automation of various sectors, industries and businesses, and they found some commonly used reports e.g. in Financial Organizations out of 10 reports 6 of them is used commonly  like yearly reports so on generic basis SAP has given us 60% of the system pre-defined
§  These contents are delivered to us in BI Content in the form of delivered versions
§  If the content satisfies your need you just have to activate it to make it available for usage
§  Otherwise you can customize it according to your needs which is termed as Enhancements

Difference between OLTP (Transactional) and OLAP (BW) Systems

§  Online Transaction Processing (OLTP) systems generates data like applications used by ATM Machines, Cashier machine in grocery shop
§  Consider a business process scenario in bank
·         Ajmal want to open an account in bank, info needed to create an account are Account ID, Name, Address, Contact No, DOB, …
·         Using Account ID assigned to Ajmal, he can perform transaction like deposits, withdrawals, fund transfers and online purchases etc.
·         This daily business process of bank generates data, as illustrated in figure on next slide

·         This daily business data is classified as Master data and Transaction data
·         Master data is a data that didn’t change for a long period of time e.g. Account ID, Name, Address or Contact No
·         Transaction data is data that change frequently e.g. withdrawal made by ATM or deposit made by PC of bank cashier
§  The sole purpose of OLTP system is that business process must not stop, it must work efficiently and must be available for 24x7
§  Reports are usually made on monthly, quarterly, half yearly or yearly bases
§  If we generate reports on huge data using transfer rules, calculation etc. will put load on OLTP systems
§  If due to this load transactions stop, i.e.  if cashier is not able to deposit money, means loss in business
§  Therefore; to avoid performance bottlenecks in OLTP systems this data will be transferred to OLAP (BI) system
§  During transfer this data is cleansed or harmonized using transfer rules for better read access
§  Thus OLAP system can’t generate data, it rely on source system (OLTP) for data

Benefits of OLAP System

§  Query performance is increased as OLTP system is free to perform transactions
§  Historical reports are possible on OLAP systems; because performance of OLTP applications depends on size of database, to increase performance of OLTP system organization only maintain 1 month, 6 months or 1 year data on OLTP system, while the old data is archived in BI system for historical reporting
§  Analytical reports
·         Analytical reports involves drill downs which effects performance of OLTP systems
·         E.g. consider a company has business in Peshawar, Lahore and Karachi, after checking the sale of Pakistan report, it’s observed that Peshawar is performing low business, to see further details of Peshawar, the user will drill down Peshawar for more information
§  BI contents allow developer not to start from scratch
·         60% system is already defined in the form BI contents
·         In BI contents  small changes can be made to objects (InfoProviders, DataSource etc.) through enhancements which reduces development time
§  Centralized repository
·         Refer to the business example in point 3 (analytical reports) if this business don’t have OLAP system on top of three individual OLTP systems, business analyst will have to approach three individual databases for reporting, thus OLAP system acts as single point of contact for an organization reporting need
§  BEx independent reporting tool, with a basic training to user a user can develop his own simple reports

Components of SAP NetWeaver BW System

§  Figure shows the components of SAP NW BW system

§  Source System
·         Source system is system which provide data to BW system
·         Figure illustrates different type of source systems and their connection types supported by SAP NW BW
·         DB Connect connection type is used to connect to relational data sources
·         UD Connect type connection is used to connect to multi-dimensional data sources as well as relational data sources
·         BI Service API is used to connect to SAP source systems like SAP ERP, SAP CRM etc.
·         File Interface is  used to connect to flat file source system like csv file
·         Web Service is used to connect to online systems e.g. to access order details from online store
·         BAPI is used for legacy system, before NetWeaver 2007 we were using  3rd party ETL tools for connection, after NW 2007 with the help of ABAPers using BAPI we extract data from legacy system
·         These connections are maintained by admin modules i.e. BASIS team
·         95% of the time you will come across flat file and sap source systems

§  Data Source (PSA)

·         DataSource(PSA) a first layer in BW system which stage the raw data from source system
·         DataSources are BI objects used to extract and stage data from source systems, the following
·         Persistent Staging Area (PSA) is transparent database table in which requested data is stored

§  Depending upon how you want to store data e.g. aggregated, detail level and what type of data you are loading e.g. transactional or master you select different data targets like ODS, VirtualProvider, InfoObjects etc.) which are termed as InfoProviders
§  Metadata and documents on the left store data about your development status e.g. objects you developed its states etc.
§  Open hub services allows to load data from BW system to other systems
§  To pass data from DataSource to a specific InfoProvider we perform a transformation i.e. implementing business logic, calculations etc.
§  Once data is available in InfoProvider we can use the analytical services like reporting etc.

Data Flow / Extract Transform and Load (ETL)

§  The ETL process or data flow is a list of steps that raw (source) data must follow to be extracted, transformed, and loaded into targets in the BI system
§  The figure on next page illustrates the data flow of BW 7.x version

§  InfoPackage
·         It is the scheduler object to load raw data from source system to DataSource (PSA)
·         InfoPackage also defines selection filters, parallel processing, date of extraction etc.
§  Transformation
·         Transformation is set of rules based on which data be loaded

§  Data Transfer Process (DTP)
·         After the InfoPackage has extracted and saved the data into PSA, you must execute DTP
·         DTP reads data from PSA or another data target, transform it, and supplies the cleansed data to the appropriate data target

§  Process Chains
·         To extract data from source system and load it into data target on regular basis, SAP BW offers a tool called Process Chains that automates this activity

No comments:

Post a Comment